From 3c44cfd495ce041769915e490a17f5cb9a6d7596 Mon Sep 17 00:00:00 2001 From: Patrick Gratz Date: Mon, 12 Jan 2026 12:17:48 +0100 Subject: [PATCH 1/2] fixed merge conflict --- Makefile | 2 +- .../local_digital_twins/index.md | 115 +++++++++--------- docs/toolbox/ldt_toolkit.md | 33 ++--- 3 files changed, 74 insertions(+), 76 deletions(-) diff --git a/Makefile b/Makefile index 72ea582d..679768fa 100644 --- a/Makefile +++ b/Makefile @@ -10,7 +10,7 @@ PYTHON := python3 .PHONY: virtenv_create ## Create virtualenv virtenv_create: - @python3 -c "import virtualenv" >/dev/null 2>&1 || pip install --break-system-packages --user virtualenv + @python3 -c "import virtualenv" >/dev/null 2>&1 || pip3 install --break-system-packages --user virtualenv python3 -m virtualenv $(VENV_NAME) source $(VENV_NAME)/bin/activate && pip install -r requirements.txt diff --git a/docs/documentation/local_digital_twins/index.md b/docs/documentation/local_digital_twins/index.md index 28375db3..7ec34f60 100644 --- a/docs/documentation/local_digital_twins/index.md +++ b/docs/documentation/local_digital_twins/index.md @@ -1,10 +1,11 @@ --- title: Local Digital Twins --- +# Local Digital Twins +Local Digital Twins (LDTs) are at the forefront of transforming how cities and communities leverage data and technology to address complex challenges. -# Towards AI-ready & interoperable LDTs - -Local Digital Twins (LDTs) are at the forefront of transforming how cities and communities leverage data and technology to address complex challenges. As the demand for AI-ready and interoperable systems grows, it becomes imperative to establish a unified approach that ensures seamless integration and collaboration across diverse LDT environments. This section introduces the collective efforts and strategic vision aimed at achieving this goal, setting the stage for the technical and strategic objectives outlined below. By fostering interoperability and aligning with open standards, we pave the way for scalable, sustainable, and innovative solutions that benefit both local and European-wide initiatives. +## Towards AI-ready & interoperable LDTs + As the demand for AI-ready and interoperable systems grows, it becomes imperative to establish a unified approach that ensures seamless integration and collaboration across diverse LDT environments. This section introduces the collective efforts and strategic vision aimed at achieving this goal, setting the stage for the technical and strategic objectives outlined below. By fostering interoperability and aligning with open standards, we pave the way for scalable, sustainable, and innovative solutions that benefit both local and European-wide initiatives. ### Strategic Goals - Assess Cross-Site Interoperability: Evaluate the feasibility and value of enhancing interoperability among LDTs across TEF sites @@ -21,7 +22,7 @@ Local Digital Twins (LDTs) are at the forefront of transforming how cities and c To facilitate collaboration and drive progress towards these goals, we established the LDT Club—a dedicated forum where stakeholders can share insights, align strategies, and collectively address challenges. This collaborative platform serves as a cornerstone for fostering innovation and ensuring that the transition to AI-ready and interoperable LDTs is both inclusive and effective. -Meeting minutes, recordings and further materials related to the LDT Club are available under [Club Meetings Sharepoint](https://imecinternational.sharepoint.com/:f:/r/sites/Citcom.aiTEF/Shared%20Documents/WP3%20-%20Infrastructure/T3.2/LDT%20Club?csf=1&web=1&e=zyhvcL). +Meeting minutes, recordings and further materials related to the LDT Club are available under . For more details on the tools and methodologies supporting the development of Local Digital Twins, refer to the [LDT Toolkit](../../toolbox/ldt_toolkit.md). @@ -29,16 +30,11 @@ For more details on the tools and methodologies supporting the development of Lo ![Scenario](img/scenario.svg) -The scenario illustrated in the diagram above puts the LDT CitiVerse EDIC and CitCom.AI TEF sites in relation. While the EDIC aims at creating a federated network of Local Digital Twins (LDTs) across Europe, CitCom.AI leverages LDTs to test and integrate AI solutions, fostering innovation in Smart Cities and Communities. - -AI innovators begin by collaborating with CitCom.AI TEF sites, where they can access real-world LDT environments to develop, test, and validate their AI solutions. Ideally, these environments should provide standardized, interoperable data and infrastructure aligned with EU LDT Toolbox guidelines. - -Once solutions are successfully demonstrated within CitCom.AI, innovators can leverage the established interoperability and best practices to integrate their solutions into the broader EDIC federated LDT network. This pathway ensures that innovations are not only tested in realistic scenarios but are also ready for seamless adoption and scaling across multiple European cities and communities through the EDIC framework. - -To ensure interoperability, the CitCom.AI LDTs, currently a set of heterogeneous and independent systems, should adopt common standards and frameworks outlined in the EU LDT Toolbox. This includes aligning data models, APIs, and communication protocols to enable seamless integration and collaboration across the EDIC federated LDT network. - -Conducting a gap analysis between CitCom.AI LDTs and the EU LDT Toolbox is crucial to identify areas where alignment is needed. This analysis will help pinpoint discrepancies in data models, APIs, and communication protocols, ensuring that AI solutions tested within CitComAI can transition seamlessly to the EDIC federated network. By addressing these gaps, the pathway for AI innovators to scale their solutions across Europe becomes more efficient and less fragmented. - +The scenario illustrated in the diagram above puts the LDT CitiVerse EDIC and CitCom.AI TEF sites in relation. While the EDIC aims at creating a federated network of Local Digital Twins (LDTs) across Europe, CitCom.AI leverages LDTs to test and integrate AI solutions, fostering innovation in Smart Cities and Communities.
+AI innovators begin by collaborating with CitCom.AI TEF sites, where they can access real-world LDT environments to develop, test, and validate their AI solutions. Ideally, these environments should provide standardized, interoperable data and infrastructure aligned with EU LDT Toolbox guidelines.
+Once solutions are successfully demonstrated within CitCom.AI, innovators can leverage the established interoperability and best practices to integrate their solutions into the broader EDIC federated LDT network. This pathway ensures that innovations are not only tested in realistic scenarios but are also ready for seamless adoption and scaling across multiple European cities and communities through the EDIC framework.
+To ensure interoperability, the CitCom.AI LDTs, currently a set of heterogeneous and independent systems, should adopt common standards and frameworks outlined in the EU LDT Toolbox. This includes aligning data models, APIs, and communication protocols to enable seamless integration and collaboration across the EDIC federated LDT network.
+Conducting a gap analysis between CitCom.AI LDTs and the EU LDT Toolbox is crucial to identify areas where alignment is needed. This analysis will help pinpoint discrepancies in data models, APIs, and communication protocols, ensuring that AI solutions tested within CitComAI can transition seamlessly to the EDIC federated network. By addressing these gaps, the pathway for AI innovators to scale their solutions across Europe becomes more efficient and less fragmented.
## How to start with a gap analysis? @@ -49,14 +45,13 @@ The diagram above illustrates the different layers and tools of the EU LDT Toolb Starting from the bottom to the top layer, the following questions help structure your gap analysis. Each question is paired with its rationale and expected outcome to guide your assessment: ### 1. Question: What are the city-services supported by your LDT? -**Related layer:** Services - -**Why it matters:** It clarifies which service domains (e.g. mobility, energy, waste) each LDT addresses. This helps identify overlaps, complementary areas, and opportunities for aligning use-case-specific models or indicators. \ +**Related layer:** Services
+**Why it matters:** It clarifies which service domains (e.g. mobility, energy, waste) each LDT addresses. This helps identify overlaps, complementary areas, and opportunities for aligning use-case-specific models or indicators.
**Outcome:** Provides a shared view of functional scope across partners, guiding discussions on domain priorities and potential areas for interoperability or joint development. -### 2. Question: What data sources do we have and in which formats?\ -**Related layer:** Data Sources \ -**Why it matters:** Data heterogeneity is a major barrier to interoperability. Knowing the sources (IoT, legacy systems, GIS, etc.) helps assess compatibility. \ +### 2. Question: What data sources do we have and in which formats?
+**Related layer:** Data Sources
+**Why it matters:** Data heterogeneity is a major barrier to interoperability. Knowing the sources (IoT, legacy systems, GIS, etc.) helps assess compatibility.
**Outcome:** Supports data model alignment, semantic mapping, and standardization efforts across partners. ### 3. Question: What standards in terms of interoperability are you already supporting? @@ -65,7 +60,7 @@ Starting from the bottom to the top layer, the following questions help structur **Outcome:** Helps to map the interoperability landscape, uncover gaps, and guide toward converging on shared mechanisms (e.g., FIWARE, IDS, GAIA-X standards). ### 4. Question: How do you retrieve, store and publish data? -**Related layer:** Data Acquisition \ +**Related layer:** Data Acquisition
**Why it matters:** Covers the full data lifecycle and reveals differences in how data is made accessible, which impacts interoperability and reuse. **Outcome:** Identifies common patterns and gaps in data handling to support alignment on shared access and publishing approaches. @@ -89,45 +84,45 @@ The following sections provide two examples where the template has been applied. ### Luxembourg Institute of Science and Technology / Luxembourg TEF - LDT For Electromobility​ -**Question:**\ -What are the city-services supported by your LDT?\ -**Answer:**\ +**Question:**
+What are the city-services supported by your LDT?
+**Answer:**
Our LDT is meant to enable city planners & mobility operators to monitor and optimise EV related infrastructure. It is supposed to empower decision-making through predictive analytics, scenario testing ("What-if scenarios"), and monitoring, fostering greener and smarter urban​ development. -**Question:** \ -What data sources do we have and in which formats?\ -**Anwer:**\ +**Question:**
+What data sources do we have and in which formats?
+**Anwer:**
We deal with data about energy consumption and production. The data comes from simulation engines, REST APIs, or historical datasets. The data format is mainly JSON. -**Question:**\ -What standards in terms of interoperability are you already supporting?\ -**Answer:**\ +**Question:**
+What standards in terms of interoperability are you already supporting?
+**Answer:**
Our architecture is based on custom micro services and Azure components. Communication happens mainly via REST APIs (documented using OpenAPI 3.0) or Azure Event Grid. Since we are relying on Azure DT our digital twin model is based on DTDL v3. For any geospatial data we rely on GeoJSON. -**Question**\ -How do you retrieve, store and publish data?\ -**Answer:**\ +**Question**
+How do you retrieve, store and publish data?
+**Answer:**
Entities and their relationships are stored in a Neo4j graph database, while telemetry data is kept as timeseries in InfluxDB. We also use Minio (S3) as a kind of staging storage. However, this data is not yet published. -**Question:**\ +**Question:**
Elaborate whether you use or plan to use AI in your LDT. If yes, what kind of AI (LLMs, predictive models, classifiers etc.) and for what purpose? -**Answer:**\ +**Answer:**
A user can define scenario/new entities parameters to trigger the generation of synthetic (simulations) or forecasting (deep learning) data. Several purposes: what-if scenarios, forecasting, anomalies detection etc. Results stored in MinIO or InfluxDB and then used for new graphs/entities. -**Questions:**\ -What data pipelines are you currently using between the different layers of your LDT?\ -**Answer:**\ +**Questions:**
+What data pipelines are you currently using between the different layers of your LDT?
+**Answer:**
We have various data pipelines that transform e.g. source data into graph entities and telemetry. Those pipelines are YAML declarations, interpreter and executed by RedPanda Connect (formerly benthos). A dedicated micro service allows to us orchestrate such pipelines. -**Question:**\ -What kind of visualizations is your LDT supporting?\ -**Answer:**\ +**Question:**
+What kind of visualizations is your LDT supporting?
+**Answer:**
All visualizations are integrated in a single LDT front-end. A mapbox based visualization shows the relevant entities (buildings with PVs, charging stations and POIs) on a geographical map. Chart.js based graphs are built according to city's needs, they can be grouped into a dashboard. Separation of real/historical telemetries and generated one. ### Aarhus City Lab / Denmark TEF - BIPED Digital Twin​ -**Question:** What are the city-services supported by your LDT?\ -**Answer:**\ +**Question:** What are the city-services supported by your LDT?
+**Answer:**
The BIPED LDT is designed to support city services primarily focused on enabling Positive Energy Districts (PEDs) and enhancing urban sustainability: - PED Planning and Development: The digital twin aims to assist city governments, @@ -149,9 +144,9 @@ Besides these defined city-services BiPED also has achieved to promote Open Data investigated how open data can bring secondary value. Also a hope for BiPED is that it can leverage the agenda for local data management in Aarhus Municipality. -**Question:** \ -What data sources do we have and in which formats?\ -**Answer:**\ +**Question:**
+What data sources do we have and in which formats?
+**Answer:**
BIPED utilizes a variety of data sources in different formats: - Hard and Soft Data: This includes sensor data, IoT data, energy consumption/production data ("hard data"), and information gathered from workshops @@ -163,7 +158,7 @@ data is also used. (D1.3 sec 3.3) 3.3) - Mobility Data: Mobility data comes from Road Twin Software. Traffic data is derived from OpenStreetMap and is accessible in a schema compatible with the INSPIRE -Transport Network. TomTom Traffic Stats data is also utilized for visualizations. D1.3 +Transport Network. TomTom Traffic Stats data is also utilized for visualizations. (D1.3 sec 3.3, D2.1 annex 2, D 2.2 table 3) - Geographic and Urban Data: This encompasses geographic data, building data, sensor data, and IoT data. Open Data from the Aarhus open data portal is a source, as @@ -173,11 +168,11 @@ Kredsløb. Weather forecasts and observations from the Danish meteorological Institute. Historical weather forecasts from the Norwegian meteorological institute. Full list: -https://docs.google.com/spreadsheets/d/1dpK9yo6ZPDXqMnP3U0ce4DHcQ6XJq2AIpIWSks4Uk38/edit?gid=185146452#gid=185146452 + -**Question:** \ -What standards in terms of interoperability are you already supporting?\ -**Answer:**\ +**Question:**
+What standards in terms of interoperability are you already supporting?
+**Answer:**
BIPED places a strong emphasis on interoperability to ensure scalable and sustainable solutions: - Minimum Interoperability Standards (MIMs): The project aims to foster an open @@ -202,9 +197,9 @@ between different components and systems. (D1.3 sec 3.4 and D2.1 sec 3.2) the INSPIRE Transport Network, further ensuring interoperability within a broader European framework. (D2.1 annex 2) -**Question:** \ -Elaborate whether you use or plan to use AI in your LDT. If yes, what kind of AI (LLMs, predictive models, classifiers etc.) and for what purpose?\ -**Answer:**\ +**Question:**
+Elaborate whether you use or plan to use AI in your LDT. If yes, what kind of AI (LLMs, predictive models, classifiers etc.) and for what purpose?
+**Answer:**
Yes, BIPED explicitly plans to use Artificial Intelligence (AI) in its Local Digital Twin (LDT). The project envisions "Future-Ready Intelligent Twins" that will leverage AI to: - Learn and make real-time accurate insights and predictions. @@ -217,9 +212,9 @@ primary AI models in BiPED currently. The project also considers AI from a regulatory perspective, referencing the "AI Act" in its Data Management Plan, suggesting a structured approach to AI integration. -**Question:** \ -What data pipelines are you currently using between the different layers of your LDT?\ -**Answer:**\ +**Question:**
+What data pipelines are you currently using between the different layers of your LDT?
+**Answer:**
The BIPED Digital Twin's architecture involves several components that interact to form data pipelines: - Architectural Components: The system is composed of back-end components, @@ -242,9 +237,9 @@ data exchange and pipeline orchestration. (D1.3 sec 3.3 and D2.1 sec 3.1). Other datasources are expected in near future, meaning the list of protocols might be extended. -**Question:** \ -What kind of visualizations is your LDT supporting?\ -**Answer:**\ +**Question:**
+What kind of visualizations is your LDT supporting?
+**Answer:**
The BIPED LDT supports various visualizations to present urban data and model outputs: - Front-End Components: The LDT utilizes "Dashboards" and a "VC Map" as primary front-end components for visualization (D2.1 sec 3.4). diff --git a/docs/toolbox/ldt_toolkit.md b/docs/toolbox/ldt_toolkit.md index 01f5d5cf..af531b7c 100644 --- a/docs/toolbox/ldt_toolkit.md +++ b/docs/toolbox/ldt_toolkit.md @@ -2,10 +2,12 @@ title: LDT Toolkit --- -# Local Digital Twin Demonstrator +# LDT Toolkit -## Introduction +## LDT Demonstrator +The LDT Toolkit is not merely a collection of tools but a deployable demonstrator that integrates various open-source components to showcase advancements in interoperability, AI-readiness, and the practical implementation of Local Digital Twins (LDTs) in real-world scenarios. +### Introduction In the context of task 3.2, and more precisely related to the gap analysis performed by the Luxembourg TEF site, an evolution of its electromobility Local Digital Twin (LDT) is planned. This effort will culminate in the development of an LDT demonstrator which will be made available to showcase advancements in interoperability and AI-readiness. The development process will follow several iterations, each aimed at progressively extending the digital twin's capabilities. These iterations align with the three twin models described in the section "Project Iterations & Twin Capabilities": the Descriptive Twin, the Predictive Twin, and the Prospective Twin. Each model builds upon the previous one, enhancing the ability to monitor, predict, and simulate real-world assets effectively. @@ -14,23 +16,23 @@ The LDT demonstrator will be built using open-source components, ensuring transp --- -## Features +### Features -### Real-Time Monitoring +#### Real-Time Monitoring - View live asset status, locations, and entity charts via the Mapbox React app. - [Access Mapbox React App]() -### Historical Dashboards +#### Historical Dashboards - Analyze historical data trends and insights using Grafana dashboards. - [Access Grafana Dashboards]() -### Analytical Tool +#### Analytical Tool - Perform advanced analytics and reporting via Superset. - [Access Superset Analytics]() --- -## Developer Resources +### Developer Resources - **[Context Broker API Documentation](https://stellio.readthedocs.io/en/latest/)**: Access the NGSI-LD API documentation and OpenAPI spec for the Stellio Context Broker. - **[Smart Data Models](https://smartdatamodels.org/)**: Browse the Smart Data Models library for standardized context information schemas. @@ -38,7 +40,7 @@ The LDT demonstrator will be built using open-source components, ensuring transp --- -## System Architecture +### System Architecture ![System Architecture Diagram](img/ldt-demonstrator.svg) @@ -46,9 +48,9 @@ The demonstrator integrates S3 and IoT data via Kafka, processes events through --- -## Development Stack +### Technology Stack -### Key Technologies +#### Key Components - **[Stellio Context Broker](https://stellio.readthedocs.io/en/latest/)**: NGSI-LD context management for smart cities. - **[NGSI-LD](https://ngsi-ld.org)**: Next Generation Service Interfaces - Linked Data. @@ -60,9 +62,9 @@ The demonstrator integrates S3 and IoT data via Kafka, processes events through --- -## Project Iterations & Digital Twin Capabilities* +### Project Iterations & Digital Twin Capabilities* -### Iteration 1: Descriptive Twin +#### Iteration 1: Descriptive Twin - Presents the current and historical state of the real-world asset, including both static and dynamic characteristics.
@@ -72,7 +74,7 @@ The demonstrator integrates S3 and IoT data via Kafka, processes events through
-### Iteration 2: Predictive Twin +#### Iteration 2: Predictive Twin - Builds on the descriptive twin by providing predictions of how the asset may evolve.
@@ -82,10 +84,10 @@ The demonstrator integrates S3 and IoT data via Kafka, processes events through
-### Iteration 3: Prospective Twin +#### Iteration 3: Prospective Twin - Enables “what-if” analysis by allowing users to simulate the impact of potential actions on the asset.
- +
Figure 3: Prospective Twin Model (Source: [ETSI GR CIM 017 V1.1.1](https://www.etsi.org/deliver/etsi_gr/CIM/001_099/017/01.01.01_60/gr_CIM017v010101p.pdf)) @@ -93,3 +95,4 @@ The demonstrator integrates S3 and IoT data via Kafka, processes events through
For further information we refer to [ETSI GR CIM 017 V1.1.1](https://www.etsi.org/deliver/etsi_gr/CIM/001_099/017/01.01.01_60/gr_CIM017v010101p.pdf): *Context Information Management (CIM); Feasibility of NGSI-LD for Digital Twins*. + From 8335759f322240004f24f578879adf1393df346b Mon Sep 17 00:00:00 2001 From: Patrick Gratz Date: Mon, 12 Jan 2026 12:22:41 +0100 Subject: [PATCH 2/2] fixed line breaks & broken references; aligned section titles with navigation entries --- Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Makefile b/Makefile index 679768fa..72ea582d 100644 --- a/Makefile +++ b/Makefile @@ -10,7 +10,7 @@ PYTHON := python3 .PHONY: virtenv_create ## Create virtualenv virtenv_create: - @python3 -c "import virtualenv" >/dev/null 2>&1 || pip3 install --break-system-packages --user virtualenv + @python3 -c "import virtualenv" >/dev/null 2>&1 || pip install --break-system-packages --user virtualenv python3 -m virtualenv $(VENV_NAME) source $(VENV_NAME)/bin/activate && pip install -r requirements.txt